Issue #1797 fix api generate index array (#1798)

Fixes #1797

# Description
This PR changes the following:

- Fix so that public API method ``GridData.generate_index_array``
returns the same as it did in iMOD Python 1.0, throw DeprecationWarning
urging users to update to ``GridData.generate_isactive_svat_array``.
- Further improve new name method to
``GridData.generate_isactive_svat_array`` and add it to public API docs
- Refactor methods in GridData somewhat, to reduce code duplication.
- Start renaming the unfortunately chosen variable name ``index`` to the
more descriptive ``isactive``, as the variable is in fact NOT an index.
In a further PR we can rename this variable throughout the complete
MetaSWAP module, which will affect a lot of files. I think that is
worthy a separate PR.

# Checklist
<!---
Before requesting review, please go through this checklist:
-->

- [x] Links to correct issue
- [x] Update changelog, if changes affect users
- [x] PR title starts with ``Issue #nr``, e.g. ``Issue #737``
- [ ] Unit tests were added
- [ ] **If feature added**: Added/extended example
- [ ] **If feature added**: Added feature to API documentation
- [ ] **If pixi.lock was changed**: Ran `pixi run generate-sbom` and
committed changes

Author: JoerivanEngelen

docs/api/changelog.rst

@@ -64,6 +64,8 @@ Changed
   :class:`imod.mf6.Evapotranspiration` are now optional variables. If provided,
   now require ``"segment"`` dimension when ``proportion_depth`` and
   ``proportion_rate``.
+- :meth:`imod.msw.GridData.generate_index_array` is now deprecated, use
+  :meth:`imod.msw.GridData.generate_isactive_svat_arrays` instead.
 
 
 [1.0.0] - 2025-11-11

docs/api/msw.rst

@@ -28,6 +28,7 @@ Grid packages
     GridData.from_imod5_data
     GridData.get_regrid_methods
     GridData.generate_index_array
+    GridData.generate_isactive_svat_arrays
     GridData.write
     Infiltration
     Infiltration.regrid_like

imod/msw/grid_data.py

@@ -1,3 +1,5 @@
+import warnings
+
 import numpy as np
 import xarray as xr
 
@@ -83,15 +85,14 @@ def __init__(
 
         self._pkgcheck()
 
-    def generate_index_array(self) -> np.ndarray:
+    def _generate_isactive_array(self) -> xr.DataArray:
         """
-        Generate index array to be used on other packages.
+        Generate a 1D array of active cells to be used on other packages.
 
         Returns
         -------
         np.ndarray
-            Index array and svat grid.
-            The index array is a 1D array with the index of the active cells.
+            A 1D array with which cells are active.
         """
         area = self.dataset["area"]
         active = self.dataset["active"]
@@ -101,36 +102,47 @@ def generate_index_array(self) -> np.ndarray:
         # https://github.com/dask/dask/issues/11753
         isactive.load()
 
-        index = isactive.values.ravel()
-
-        return index
+        return isactive
 
-    def generate_index_svat_array(self) -> tuple[np.ndarray, xr.DataArray]:
+    def generate_isactive_svat_arrays(self) -> tuple[np.ndarray, xr.DataArray]:
         """
         Generate index array and svat grid to be used on other packages.
 
         Returns
         -------
         tuple[np.ndarray, xr.DataArray]
-            Index array and svat grid.
-            The index array is a 1D array with the index of the active cells.
+            isactive array and svat grid.
+            The isactive array is a 1D array with which cells are active.
             The svat grid is a 2D array with the SVAT numbers for each cell.
         """
-        index = self.generate_index_array()
+        isactive = self._generate_isactive_array()
+        isactive_1d = isactive.values.ravel()
 
-        area = self.dataset["area"]
-        active = self.dataset["active"]
-        isactive = area.where(active).notnull()
+        svat = xr.full_like(isactive, fill_value=0, dtype=np.int64).rename("svat")
+        svat.data[isactive.data] = np.arange(1, isactive_1d.sum() + 1)
 
-        svat = xr.full_like(area, fill_value=0, dtype=np.int64).rename("svat")
-        # Load into memory to avoid dask issue
-        # https://github.com/dask/dask/issues/11753
-        isactive.load()
-        svat.load()
+        return isactive_1d, svat
 
-        svat.data[isactive.data] = np.arange(1, index.sum() + 1)
+    def generate_index_array(self) -> tuple[np.ndarray, xr.DataArray]:
+        """
+        This method is kept for backward compatibility, but will be removed in
+        future versions and will thus throw a deprecation warning. Use
+        :meth:`imod.msw.GridData.generate_isactive_svat_arrays` instead.
+
+        Generate index array and svat grid to be used on other packages.
 
-        return index, svat
+        Returns
+        -------
+        tuple[np.ndarray, xr.DataArray]
+            isactive array and svat grid.
+            The isactive array is a 1D array with which cells are active.
+            The svat grid is a 2D array with the SVAT numbers for each cell.
+        """
+        warnings.warn(
+            "Method 'generate_index_array' is deprecated and will be removed in the future, use 'generate_isactive_svat_arrays' instead.",
+            DeprecationWarning,
+        )
+        return self.generate_isactive_svat_arrays()
 
     def _pkgcheck(self):
         super()._pkgcheck()

imod/msw/model.py

@@ -360,7 +360,7 @@ def write(
         # Get index and svat
         grid_key = self.get_pkgkey(GridData)
         grid_pkg = cast(GridData, self[grid_key])
-        index, svat = grid_pkg.generate_index_svat_array()
+        index, svat = grid_pkg.generate_isactive_svat_arrays()
 
         # write package contents
         for pkgname in self:
@@ -565,11 +565,13 @@ def split(
 
         for submodel_name, submodel in partitioned_submodels.items():
             partition_info = submodel_to_partition[submodel_name]
-            sliced_grid_pkg = clip_by_grid(grid_pkg, partition_info.active_domain)
-            sliced_index = sliced_grid_pkg.generate_index_array()
+            sliced_grid_pkg = cast(
+                GridData, clip_by_grid(grid_pkg, partition_info.active_domain)
+            )
+            sliced_isactive = sliced_grid_pkg._generate_isactive_array().values
 
             # Add package to model if it has data in the active domain.
-            if bool(sliced_index.any()):
+            if bool(sliced_isactive.any()):
                 is_in_active_domain[submodel_name] = True
                 submodel[grid_key] = sliced_grid_pkg
             else:

imod/tests/test_msw/test_grid_data.py

@@ -64,7 +64,7 @@ def test_write(
         xr.full_like(like, True, dtype=bool),
     )
 
-    index, svat = grid_data.generate_index_svat_array()
+    index, svat = grid_data.generate_isactive_svat_arrays()
 
     with tempfile.TemporaryDirectory() as output_dir:
         output_dir = Path(output_dir)
@@ -293,12 +293,12 @@ def case_grid_data_two_subunits__dask(
 
 
 @parametrize_with_cases("grid_data_dict", cases=".", has_tag="two_subunit")
-def test_generate_index_svat_array(
+def test_generate_isactive_svat_arrays(
     grid_data_dict: dict[str, xr.DataArray], coords_two_subunit: dict
 ):
     grid_data = GridData(**grid_data_dict)
 
-    index, svat = grid_data.generate_index_svat_array()
+    index, svat = grid_data.generate_isactive_svat_arrays()
 
     index_expected = [
         False,
@@ -346,7 +346,7 @@ def test_generate_index_svat_array(
 def test_simple_model(fixed_format_parser, grid_data_dict: dict[str, xr.DataArray]):
     grid_data = GridData(**grid_data_dict)
 
-    index, svat = grid_data.generate_index_svat_array()
+    index, svat = grid_data.generate_isactive_svat_arrays()
 
     with tempfile.TemporaryDirectory() as output_dir:
         output_dir = Path(output_dir)
@@ -388,7 +388,7 @@ def test_simple_model_1_subunit(
 ):
     grid_data = GridData(**grid_data_dict)
 
-    index, svat = grid_data.generate_index_svat_array()
+    index, svat = grid_data.generate_isactive_svat_arrays()
 
     with tempfile.TemporaryDirectory() as output_dir:
         output_dir = Path(output_dir)